On Sep 4, 2006, at 1:35 AM, kiersb@xs4all.net wrote:

>> Number:         34461
>> Category:       kern
>> Synopsis:       multiple problems; ioapic related?
>> Confidential:   no
>> Severity:       critical
>> Priority:       high
>> Responsible:    kern-bug-people
>> State:          open
>> Class:          sw-bug
>> Submitter-Id:   net
>> Arrival-Date:   Mon Sep 04 00:35:00 +0000 2006
>> Originator:     kiersb@xs4all.net
>> Release:        NetBSD 4.99.1
>> Organization:
> 	XS4All
>> Environment:
> 	
> 	
> System: NetBSD raidtest2 4.99.1 NetBSD 4.99.1 (GENERIC.MPACPI) #0:  
> Wed Aug 16 17:48:50 CEST 2006 kiers@kleurtjes:/disk2/obj/sys/arch/ 
> i386/compile/GENERIC.MPACPI i386
> Architecture: i386
> Machine: i386
>> Description:
> 	
> While rsyncing 1.3 TB from an ufs file system to an lfs file system  
> on same box, after around 200 MB:
>
> db{0}> bt
> Xintr_lapic_ipi() at netbsd:Xintr_lapic_ipi+0x7
> --- interrupt ---
> Bad frame pointer: 0xc3077318
> 0x30:
>
> then, after rebooting and issueing save rsync command, after 392 GB:
>
> sd5(isp0:0:21:0): unable to allocate scsipi_xfer
>
> and nothing special in kernel messages; after this sd5 message, the  
> computer is *very* slow,
>
> root@raidtest2:~# uptime
> 12:58AM  up  9:11, 2 users, load averages: 8.66, 8.59, 8.52
> took around two minutes. Even typing 'uptime' on the serial console  
> echoes around 1 char in two seconds.

Can you monitor disk I/O with iostat at the same time you are doing a  
copy ?

>
> dmesg available on http://dia.zepam.nl/pr-20060904-1-dmesg.txt
>
> Not production system, so I can test stuff.
>> How-To-Repeat:
> 	
>> Fix:
> 	
>
>> Unformatted:
>  	
>  	



	-- Rui Paulo